Tuesday, October 6, 2020

More Data on the Unbelievable Explosion of 81408 Spending at Medicare: Novitas and FCSO MACs Only, $300M Lost

 A few weeks ago, CMS released state and national level data on Medicare Part B spending by CPT code for CY2019.  

I ran an analysis on MoPath spending and found that 81408 -  the Tier 2, Level 9 CPT code for rare fully sequenced genes - had exploded from 2017 to 2018 to 2019.   This explosion was concurrent with DOJ fraud cases alleging billions of dollars in fraudulent MoPath charges.   See an entry point here for the DOJ story and links therein.  See my original article on CY2019 Medicare data here.

From previous blogs in this series, chiefly here, we knew that 81408 spending exploded from almost nothing in 2017 to 2018 to 2019.  It rose from nowhere to be the highest paid CPT code in MoPath in 2019.  See figure:

See the definition of 81408 (orphan, full sequence genes) here.  

We know from the DEX (TM) public database and the MolDx edit file (obtained via FOIA from CMS) that MolDx doesn't allow any billing on 81408 for any medical purpose. (FN1) 

Where the Money Flowed; Two MACs

We now know for sure that the explosive growth in 81408 spending was nearly entirely in the Novitas and FCSO MACs (both of which are part of Guidewell, also holds Florida BCBS).   

NGS MAC paid essentially nothing for code 81408.  Neither were there payments for this code from the MolDx MACs (WPS, Noridian, CGS, Palmetto).   We also know that the 2017 spending was entirely in the now-defunct Cahaba MAC and nowhere in the U.S. except labs in GA and TN.  

Before the Deluge: 2017 Spending on 81408

2017 81408 spending was small ($9M). But when you look at it, it was already very atypical - all billed in two states that had little share of the national mopath billing, but had a monopoly on the 81408 billing.  

We know from the 2017 CMS physician/lab utilization data that only 9 entities nationwide were paid in 2017 for 81408 billing.  All labs receiving payments for 81408 in 2017 were in the now-defunct Cahaba MAC.

Major recognized labs like Invitae, Bioreference, Quest, Labcorp, etc, had $0 cases for 81408 in the 2017 CMS database.  LabSolutions was linked in a newspaper report to a DOJ genetic test fraud investigation (here).  See also here and here.  It had nearly 60% of all USA Medicare billings for 81408 in 2017 as shown in the table above.   By Fall 2019,  however, Lab Solutions had been "paid $154M" by Medicare and Medicare Advantage (here).   Since Georgia now falls under MolDx jurisdiction, we noted that as of September 2020, LabSolutions has 113 Z codes filed with Palmetto MolDx, of which 17 are classed as payable but the other 96 are listed with no payment edits.  While this information can be gleaned from the Palmetto MolDx "MEF" file, a FOIA document, the exact same data is available through the Palmetto DEX public website here.  Codes were obtained by LabSolutions across a range of times in 2018, 2019, and a few into 2020, well after the DOJ investigation was announced in Fall 2019 (see prior links in this paragraph).

Google has little data on Total Diagnostix but will give you a bankruptcy filing link.  Rounding out the top three billers for 81408, there's not much to find on National Premier Laboratories other than a trademark registration.

We know from 2019 data (a time when Palmetto held GA and TN) that spending in GA and TN for 81408 fell to $0 in 2019, from $9M in 2017.  So it's very easy to stop 81408 spending cold: Both NGS MAC and Palmetto MAC accomplished it.

What we know now is...over the next 18 months, 81408 spending would explode 32X, from $9M to $290M.

2019 MAC Spending on 81408 by MAC 

I've now assessed $290M in 2019 CMS spending on 81408 by MAC, aggregating to MACs from 2019 state spending files here.  It doesn't look good - 89% via the Novitas MAC and 11% via the FCSO MAC (Florida).   

As I've previously shown, the Novitas LCD for molecular testing publicly announces there are NO EDITS on 81408 "because it is a code for multiple biomarkers."   Bad idea.  And as shown in the 2017/2018/2019 bar chart I opened with, this problem has been going on for a LONG time, from 2017 to 2020.

CMS Could Release 2018 Data by Lab Getting the Money - 
But CMS is Holding It

CMS releases a public database of all Part B spending by provider and CPT code, and has done so since about 2014, thanks to work by ProPublica, WSJ, and court cases.  

However, this data is normally released in May, and this year, the 2018 data is locked down.   It's October, and the 2018 data isn't released yet.  We know who got the $9M of 81408 spending in 2017 (shown above) but we don't know who got the $120M in 2018, and normally we would.  My blog on this topic here.  For example, the public should be able to look up (in a normal year, beginning in May) who were the 2018 recipients of the 81408 dollars in Oklahoma, Texas, New Jersey, Pennsylvania, and see if those recipients (for example) were also big political donors.

We Also Know State-by-State 81408 Spending in 2019


Remarkably, Oklahoma had the highest payments for 81408, totaling half to that state's mopath spending (much of the rest was BRCA 81162, another easily-paid code that's also nearly $2000 per test).


This article focused only on 81408 spending at Novitas/FCSO; see a broader view of all MoPath spending in these and other MACs here.  Spending per bene was about $3 at the NGS MAC (leaving out just two nationally edited codes, Cologuard and FMI F1 CDx), and about $29 at the consortium of MolDx MACs (including many legitimate national genomic labs like Myriad, Invitae, Labcorp, etc, billing to those MACs).   Spending at Novitas and FCSO MACs was circa $50 per bene.

Example of Palmetto MolDx noncoverage of 81408; open public DEX (TM) screen shot here.  MolDx and NGS MAC took the Nancy Reagan approach to 81408 spending: just say no.


81408 is non covered by United Healthcare, here.